]> git.saurik.com Git - wxWidgets.git/blob - wxPython/distutils/util.py
Cleanup
[wxWidgets.git] / wxPython / distutils / util.py
1 """distutils.util
2
3 Miscellaneous utility functions -- anything that doesn't fit into
4 one of the other *util.py modules.
5 """
6
7 __revision__ = "$Id$"
8
9 import sys, os, string, re
10 from distutils.errors import DistutilsPlatformError
11 from distutils.dep_util import newer
12 from distutils.spawn import spawn
13 from distutils import log
14
15 def get_platform ():
16 """Return a string that identifies the current platform. This is used
17 mainly to distinguish platform-specific build directories and
18 platform-specific built distributions. Typically includes the OS name
19 and version and the architecture (as supplied by 'os.uname()'),
20 although the exact information included depends on the OS; eg. for IRIX
21 the architecture isn't particularly important (IRIX only runs on SGI
22 hardware), but for Linux the kernel version isn't particularly
23 important.
24
25 Examples of returned values:
26 linux-i586
27 linux-alpha (?)
28 solaris-2.6-sun4u
29 irix-5.3
30 irix64-6.2
31
32 For non-POSIX platforms, currently just returns 'sys.platform'.
33 """
34 if os.name != "posix" or not hasattr(os, 'uname'):
35 # XXX what about the architecture? NT is Intel or Alpha,
36 # Mac OS is M68k or PPC, etc.
37 return sys.platform
38
39 # Try to distinguish various flavours of Unix
40
41 (osname, host, release, version, machine) = os.uname()
42
43 # Convert the OS name to lowercase, remove '/' characters
44 # (to accommodate BSD/OS), and translate spaces (for "Power Macintosh")
45 osname = string.lower(osname)
46 osname = string.replace(osname, '/', '')
47 machine = string.replace(machine, ' ', '_')
48
49 if osname[:5] == "linux":
50 # At least on Linux/Intel, 'machine' is the processor --
51 # i386, etc.
52 # XXX what about Alpha, SPARC, etc?
53 return "%s-%s" % (osname, machine)
54 elif osname[:5] == "sunos":
55 if release[0] >= "5": # SunOS 5 == Solaris 2
56 osname = "solaris"
57 release = "%d.%s" % (int(release[0]) - 3, release[2:])
58 # fall through to standard osname-release-machine representation
59 elif osname[:4] == "irix": # could be "irix64"!
60 return "%s-%s" % (osname, release)
61 elif osname[:3] == "aix":
62 return "%s-%s.%s" % (osname, version, release)
63 elif osname[:6] == "cygwin":
64 osname = "cygwin"
65 rel_re = re.compile (r'[\d.]+')
66 m = rel_re.match(release)
67 if m:
68 release = m.group()
69
70 return "%s-%s-%s" % (osname, release, machine)
71
72 # get_platform ()
73
74
75 def convert_path (pathname):
76 """Return 'pathname' as a name that will work on the native filesystem,
77 i.e. split it on '/' and put it back together again using the current
78 directory separator. Needed because filenames in the setup script are
79 always supplied in Unix style, and have to be converted to the local
80 convention before we can actually use them in the filesystem. Raises
81 ValueError on non-Unix-ish systems if 'pathname' either starts or
82 ends with a slash.
83 """
84 if os.sep == '/':
85 return pathname
86 if not pathname:
87 return pathname
88 if pathname[0] == '/':
89 raise ValueError, "path '%s' cannot be absolute" % pathname
90 if pathname[-1] == '/':
91 raise ValueError, "path '%s' cannot end with '/'" % pathname
92
93 paths = string.split(pathname, '/')
94 while '.' in paths:
95 paths.remove('.')
96 if not paths:
97 return os.curdir
98 return apply(os.path.join, paths)
99
100 # convert_path ()
101
102
103 def change_root (new_root, pathname):
104 """Return 'pathname' with 'new_root' prepended. If 'pathname' is
105 relative, this is equivalent to "os.path.join(new_root,pathname)".
106 Otherwise, it requires making 'pathname' relative and then joining the
107 two, which is tricky on DOS/Windows and Mac OS.
108 """
109 if os.name == 'posix':
110 if not os.path.isabs(pathname):
111 return os.path.join(new_root, pathname)
112 else:
113 return os.path.join(new_root, pathname[1:])
114
115 elif os.name == 'nt':
116 (drive, path) = os.path.splitdrive(pathname)
117 if path[0] == '\\':
118 path = path[1:]
119 return os.path.join(new_root, path)
120
121 elif os.name == 'os2':
122 (drive, path) = os.path.splitdrive(pathname)
123 if path[0] == os.sep:
124 path = path[1:]
125 return os.path.join(new_root, path)
126
127 elif os.name == 'mac':
128 if not os.path.isabs(pathname):
129 return os.path.join(new_root, pathname)
130 else:
131 # Chop off volume name from start of path
132 elements = string.split(pathname, ":", 1)
133 pathname = ":" + elements[1]
134 return os.path.join(new_root, pathname)
135
136 else:
137 raise DistutilsPlatformError, \
138 "nothing known about platform '%s'" % os.name
139
140
141 _environ_checked = 0
142 def check_environ ():
143 """Ensure that 'os.environ' has all the environment variables we
144 guarantee that users can use in config files, command-line options,
145 etc. Currently this includes:
146 HOME - user's home directory (Unix only)
147 PLAT - description of the current platform, including hardware
148 and OS (see 'get_platform()')
149 """
150 global _environ_checked
151 if _environ_checked:
152 return
153
154 if os.name == 'posix' and not os.environ.has_key('HOME'):
155 import pwd
156 os.environ['HOME'] = pwd.getpwuid(os.getuid())[5]
157
158 if not os.environ.has_key('PLAT'):
159 os.environ['PLAT'] = get_platform()
160
161 _environ_checked = 1
162
163
164 def subst_vars (s, local_vars):
165 """Perform shell/Perl-style variable substitution on 'string'. Every
166 occurrence of '$' followed by a name is considered a variable, and
167 variable is substituted by the value found in the 'local_vars'
168 dictionary, or in 'os.environ' if it's not in 'local_vars'.
169 'os.environ' is first checked/augmented to guarantee that it contains
170 certain values: see 'check_environ()'. Raise ValueError for any
171 variables not found in either 'local_vars' or 'os.environ'.
172 """
173 check_environ()
174 def _subst (match, local_vars=local_vars):
175 var_name = match.group(1)
176 if local_vars.has_key(var_name):
177 return str(local_vars[var_name])
178 else:
179 return os.environ[var_name]
180
181 try:
182 return re.sub(r'\$([a-zA-Z_][a-zA-Z_0-9]*)', _subst, s)
183 except KeyError, var:
184 raise ValueError, "invalid variable '$%s'" % var
185
186 # subst_vars ()
187
188
189 def grok_environment_error (exc, prefix="error: "):
190 """Generate a useful error message from an EnvironmentError (IOError or
191 OSError) exception object. Handles Python 1.5.1 and 1.5.2 styles, and
192 does what it can to deal with exception objects that don't have a
193 filename (which happens when the error is due to a two-file operation,
194 such as 'rename()' or 'link()'. Returns the error message as a string
195 prefixed with 'prefix'.
196 """
197 # check for Python 1.5.2-style {IO,OS}Error exception objects
198 if hasattr(exc, 'filename') and hasattr(exc, 'strerror'):
199 if exc.filename:
200 error = prefix + "%s: %s" % (exc.filename, exc.strerror)
201 else:
202 # two-argument functions in posix module don't
203 # include the filename in the exception object!
204 error = prefix + "%s" % exc.strerror
205 else:
206 error = prefix + str(exc[-1])
207
208 return error
209
210
211 # Needed by 'split_quoted()'
212 _wordchars_re = re.compile(r'[^\\\'\"%s ]*' % string.whitespace)
213 _squote_re = re.compile(r"'(?:[^'\\]|\\.)*'")
214 _dquote_re = re.compile(r'"(?:[^"\\]|\\.)*"')
215
216 def split_quoted (s):
217 """Split a string up according to Unix shell-like rules for quotes and
218 backslashes. In short: words are delimited by spaces, as long as those
219 spaces are not escaped by a backslash, or inside a quoted string.
220 Single and double quotes are equivalent, and the quote characters can
221 be backslash-escaped. The backslash is stripped from any two-character
222 escape sequence, leaving only the escaped character. The quote
223 characters are stripped from any quoted string. Returns a list of
224 words.
225 """
226
227 # This is a nice algorithm for splitting up a single string, since it
228 # doesn't require character-by-character examination. It was a little
229 # bit of a brain-bender to get it working right, though...
230
231 s = string.strip(s)
232 words = []
233 pos = 0
234
235 while s:
236 m = _wordchars_re.match(s, pos)
237 end = m.end()
238 if end == len(s):
239 words.append(s[:end])
240 break
241
242 if s[end] in string.whitespace: # unescaped, unquoted whitespace: now
243 words.append(s[:end]) # we definitely have a word delimiter
244 s = string.lstrip(s[end:])
245 pos = 0
246
247 elif s[end] == '\\': # preserve whatever is being escaped;
248 # will become part of the current word
249 s = s[:end] + s[end+1:]
250 pos = end+1
251
252 else:
253 if s[end] == "'": # slurp singly-quoted string
254 m = _squote_re.match(s, end)
255 elif s[end] == '"': # slurp doubly-quoted string
256 m = _dquote_re.match(s, end)
257 else:
258 raise RuntimeError, \
259 "this can't happen (bad char '%c')" % s[end]
260
261 if m is None:
262 raise ValueError, \
263 "bad string (mismatched %s quotes?)" % s[end]
264
265 (beg, end) = m.span()
266 s = s[:beg] + s[beg+1:end-1] + s[end:]
267 pos = m.end() - 2
268
269 if pos >= len(s):
270 words.append(s)
271 break
272
273 return words
274
275 # split_quoted ()
276
277
278 def execute (func, args, msg=None, verbose=0, dry_run=0):
279 """Perform some action that affects the outside world (eg. by
280 writing to the filesystem). Such actions are special because they
281 are disabled by the 'dry_run' flag. This method takes care of all
282 that bureaucracy for you; all you have to do is supply the
283 function to call and an argument tuple for it (to embody the
284 "external action" being performed), and an optional message to
285 print.
286 """
287 if msg is None:
288 msg = "%s%s" % (func.__name__, `args`)
289 if msg[-2:] == ',)': # correct for singleton tuple
290 msg = msg[0:-2] + ')'
291
292 log.info(msg)
293 if not dry_run:
294 apply(func, args)
295
296
297 def strtobool (val):
298 """Convert a string representation of truth to true (1) or false (0).
299
300 True values are 'y', 'yes', 't', 'true', 'on', and '1'; false values
301 are 'n', 'no', 'f', 'false', 'off', and '0'. Raises ValueError if
302 'val' is anything else.
303 """
304 val = string.lower(val)
305 if val in ('y', 'yes', 't', 'true', 'on', '1'):
306 return 1
307 elif val in ('n', 'no', 'f', 'false', 'off', '0'):
308 return 0
309 else:
310 raise ValueError, "invalid truth value %s" % `val`
311
312
313 def byte_compile (py_files,
314 optimize=0, force=0,
315 prefix=None, base_dir=None,
316 verbose=1, dry_run=0,
317 direct=None):
318 """Byte-compile a collection of Python source files to either .pyc
319 or .pyo files in the same directory. 'py_files' is a list of files
320 to compile; any files that don't end in ".py" are silently skipped.
321 'optimize' must be one of the following:
322 0 - don't optimize (generate .pyc)
323 1 - normal optimization (like "python -O")
324 2 - extra optimization (like "python -OO")
325 If 'force' is true, all files are recompiled regardless of
326 timestamps.
327
328 The source filename encoded in each bytecode file defaults to the
329 filenames listed in 'py_files'; you can modify these with 'prefix' and
330 'basedir'. 'prefix' is a string that will be stripped off of each
331 source filename, and 'base_dir' is a directory name that will be
332 prepended (after 'prefix' is stripped). You can supply either or both
333 (or neither) of 'prefix' and 'base_dir', as you wish.
334
335 If 'dry_run' is true, doesn't actually do anything that would
336 affect the filesystem.
337
338 Byte-compilation is either done directly in this interpreter process
339 with the standard py_compile module, or indirectly by writing a
340 temporary script and executing it. Normally, you should let
341 'byte_compile()' figure out to use direct compilation or not (see
342 the source for details). The 'direct' flag is used by the script
343 generated in indirect mode; unless you know what you're doing, leave
344 it set to None.
345 """
346
347 # First, if the caller didn't force us into direct or indirect mode,
348 # figure out which mode we should be in. We take a conservative
349 # approach: choose direct mode *only* if the current interpreter is
350 # in debug mode and optimize is 0. If we're not in debug mode (-O
351 # or -OO), we don't know which level of optimization this
352 # interpreter is running with, so we can't do direct
353 # byte-compilation and be certain that it's the right thing. Thus,
354 # always compile indirectly if the current interpreter is in either
355 # optimize mode, or if either optimization level was requested by
356 # the caller.
357 if direct is None:
358 direct = (__debug__ and optimize == 0)
359
360 # "Indirect" byte-compilation: write a temporary script and then
361 # run it with the appropriate flags.
362 if not direct:
363 try:
364 from tempfile import mkstemp
365 (script_fd, script_name) = mkstemp(".py")
366 except ImportError:
367 from tempfile import mktemp
368 (script_fd, script_name) = None, mktemp(".py")
369 log.info("writing byte-compilation script '%s'", script_name)
370 if not dry_run:
371 if script_fd is not None:
372 script = os.fdopen(script_fd, "w")
373 else:
374 script = open(script_name, "w")
375
376 script.write("""\
377 from distutils.util import byte_compile
378 files = [
379 """)
380
381 # XXX would be nice to write absolute filenames, just for
382 # safety's sake (script should be more robust in the face of
383 # chdir'ing before running it). But this requires abspath'ing
384 # 'prefix' as well, and that breaks the hack in build_lib's
385 # 'byte_compile()' method that carefully tacks on a trailing
386 # slash (os.sep really) to make sure the prefix here is "just
387 # right". This whole prefix business is rather delicate -- the
388 # problem is that it's really a directory, but I'm treating it
389 # as a dumb string, so trailing slashes and so forth matter.
390
391 #py_files = map(os.path.abspath, py_files)
392 #if prefix:
393 # prefix = os.path.abspath(prefix)
394
395 script.write(string.join(map(repr, py_files), ",\n") + "]\n")
396 script.write("""
397 byte_compile(files, optimize=%s, force=%s,
398 prefix=%s, base_dir=%s,
399 verbose=%s, dry_run=0,
400 direct=1)
401 """ % (`optimize`, `force`, `prefix`, `base_dir`, `verbose`))
402
403 script.close()
404
405 cmd = [sys.executable, script_name]
406 if optimize == 1:
407 cmd.insert(1, "-O")
408 elif optimize == 2:
409 cmd.insert(1, "-OO")
410 spawn(cmd, dry_run=dry_run)
411 execute(os.remove, (script_name,), "removing %s" % script_name,
412 dry_run=dry_run)
413
414 # "Direct" byte-compilation: use the py_compile module to compile
415 # right here, right now. Note that the script generated in indirect
416 # mode simply calls 'byte_compile()' in direct mode, a weird sort of
417 # cross-process recursion. Hey, it works!
418 else:
419 from py_compile import compile
420
421 for file in py_files:
422 if file[-3:] != ".py":
423 # This lets us be lazy and not filter filenames in
424 # the "install_lib" command.
425 continue
426
427 # Terminology from the py_compile module:
428 # cfile - byte-compiled file
429 # dfile - purported source filename (same as 'file' by default)
430 cfile = file + (__debug__ and "c" or "o")
431 dfile = file
432 if prefix:
433 if file[:len(prefix)] != prefix:
434 raise ValueError, \
435 ("invalid prefix: filename %s doesn't start with %s"
436 % (`file`, `prefix`))
437 dfile = dfile[len(prefix):]
438 if base_dir:
439 dfile = os.path.join(base_dir, dfile)
440
441 cfile_base = os.path.basename(cfile)
442 if direct:
443 if force or newer(file, cfile):
444 log.info("byte-compiling %s to %s", file, cfile_base)
445 if not dry_run:
446 compile(file, cfile, dfile)
447 else:
448 log.debug("skipping byte-compilation of %s to %s",
449 file, cfile_base)
450
451 # byte_compile ()
452
453 def rfc822_escape (header):
454 """Return a version of the string escaped for inclusion in an
455 RFC-822 header, by ensuring there are 8 spaces space after each newline.
456 """
457 lines = string.split(header, '\n')
458 lines = map(string.strip, lines)
459 header = string.join(lines, '\n' + 8*' ')
460 return header